1. Technical Field
The example embodiments described herein relate to image processing, and more particularly, to an image processing device capable of generating data corresponding to a foreground object of an image using depth data and amplitude data, and an image processing method thereof.
2. Discussion of Related Art
A depth camera is a device to acquire an image expressing a real space with three-dimensional data. The depth camera acquires depth data corresponding to distance between a camera and an object. A typical method for obtaining the depth data is a Time-of-Flight (“ToF”) method.
A ToF method may include a ToF system that issues a modulated signal to an object and receives a signal reflected from the object. The ToF system measures time spent while the reflected signal is being returned, a phase shift of the reflected signal, etc. The ToF system acquires depth data corresponding to a distance between a camera and the object based on the measurement result. Data with respect to a specific object captured on an image can be extracted by using the depth data acquired by the ToF system. For example, if a background object and a foreground object are captured in an image, a distance between the camera and the foreground object is shorter than a distance between the camera and the background object. Data with respect to the foreground object can be extracted based on a difference between the distances.
Typically, the ToF system compares depth data of a captured image and depth data of a reference background. The ToF system decides that an object captured in an image area is a foreground object when the object has depth data having depth values that are different from depth values of depth data of the reference background from among depth data of the captured image. However, depth data with respect to an outline of the foreground object or depth data with respect to an image area captured without appropriately supplying light may be noise having a depth value different from a real depth value. The noise may have a depth value different from that of the reference background, which may increase a likelihood that the noise is determined to be a foreground object. That is, the noise may make it difficult to extract data of the foreground object appropriately.